The additive divisor problem and its analogs for Fourier coefficients of cusp forms. I (Q5961416)

A unified approach is developed for sums of the form \(\sum_{n\leq x} c(n)c (n+ m)\), where \(m\) is a positive integer and \(c(n)\) stands either for the usual divisor function \(d(n)\), or for the Fourier coefficients of holomorphic or non-holomorphic cusp forms. The method is based on the generating Dirichlet series \(\sum^\infty_{n=1} c(n)c(n+m) (n+m)^{-s}\), which can be analytically continued to a meromorphic function by the spectral theory. An argument like this has been previously worked out by \textit{A. Good} [Math. Z. 183, 95-129 (1983; Zbl 0497.10015)] for holomorphic cusp forms, and by \textit{L. A. Takhtadzhyan} and \textit{A. I. Vinogradov} [Zap. Nauchn. Semin. LOMI 134, 84-116 (1984; Zbl 0536.10025)] for the ordinary additive divisor problem. The remaining case pertaining to a non-holomorphic cusp form \(u(z)\) can be reduced to the study of the spectral decomposition of \(u^2(z)\), in particular to the estimation of the inner products of this function against Maass wave forms. The corresponding problem for holomorphic cusp forms was solved by \textit{A. Good} [Math. Ann. 255, 523-548 (1981; Zbl 0439.30031)], but the case of nonholomorphic cusp forms remained open. In this paper, a new variant of the Rankin-Selberg method is given which applies to all the inner products in question. Those involving holomorphic cusp forms are considered in detail in the present Part I, and Part II will be devoted to an analogous analysis for nonholomorphic cusp forms.